Natural Hazards and Earth System Sciences (Nov 2019)
Hydro-meteorological reconstruction and geomorphological impact assessment of the October 2018 catastrophic flash flood at Sant Llorenç, Mallorca (Spain)
Abstract
An extraordinary convective rainfall event, unforeseen by most numerical weather prediction models, generated a devastating flash flood (305 m3 s−1) in the town of Sant Llorenç des Cardassar, Mallorca, on 9 October 2018. Four people died inside this village, while casualties were up to 13 over the entire affected area. This extreme event has been reconstructed by implementing an integrated flash flood modelling approach in the Ses Planes catchment up to Sant Llorenç (23.4 km2), based on three components: (i) generation of radar-derived precipitation estimates, (ii) modelling of accurate discharge hydrographs yielded by the catchment (using FEST and KLEM models), and (iii) hydraulic simulation of the event and mapping of affected areas (using HEC-RAS). Radar-derived rainfall estimates show very high agreement with rain gauge data (R2=0.98). Modelled flooding extent is in close agreement with the observed extension by the Copernicus Emergency Management Service, based on Sentinel-1 imagery, and both far exceed the extension for a 500-year return period flood. Hydraulic simulation revealed that water reached a depth of 3 m at some points, and modelled water depths highly correlate (R2=0.91) with in situ after-event measurements. The 9 October flash flood eroded and transported woody and abundant sediment debris, changing channel geomorphology. Water velocity greatly increased at bridge locations crossing the river channel, especially at those closer to the Sant Llorenç town centre. This study highlights how the very low predictability of this type of extreme convective rainfall events and the very short hydrological response times typical of small Mediterranean catchments continue to challenge the implementation of early warning systems, which effectively reduce people's exposure to flash flood risk in the region.